A technical discussion of the Borresen ironless woofer.

[smowry]

The primary issues with any ironless transducer motor include but are not limited to the following.

A. Magnets' operating point and its relation to potential demag.
B. Manufacturability
C. Cost containment

The Borresen ironless woofer is shown below.

 

[smowry]

I will start this discussion with my ironless woofer concept, STEALLUS. STEALLUS is an acronym that I assigned to my ironless woofer concept, where the words STEALTH and PHALLUS are truncated to form STEALLUS. Here's a link to a two-part discussion that I published in 2005 in Voice Coil magazine, edited by Vance Dickason.

https://pearl-hifi.com/06_Lit_Archive/14_Books_Tech_Papers/Mowry_Steve/Steallus_Motor_Design.pdf

The STEALLUS transducer topology is based on a concept from William N. House for Harman International. House filed for his US Patent in March of 1991.

https://www.freepatentsonline.com/5142260.pdf

There's a lot here and if you plan to participate in this discussion, please consider reviewing the documents with links provided above.

 

[smowry]

Then I will introduce another ironless motor concept that I called Almost air core and is based on my work alone. I published a discussion again in Voice Coil magazine and again the discussion was edited by Vance Dickason.

https://pearl-hifi.com/06_Lit_Archi.../Mowry_Steve/Air_Core_Tweeter_Magnet_Assy.pdf

Please keep in mind that STEALLUS is a woofer (today it would be called a subwoofer) while Almost air core is a tweeter. This is important as we shall see later on.

 

[NTK]

I know squat about magnetics, am not even an EE, but reasonably good at blindly following examples. Interested in seeing how the magnetic field may look with just those stacked magnets. Modified the Mathematica example. This wasn't even an axisymmetric model (2-D Cartesian), just wanted a picture. Another case of garbage-in, (but to the uninitiated, impressively looking) garbage-out

MagnetostaticPDEComponent: Magnetostatic equation—Wolfram Documentation

MagnetostaticPDEComponent is typically used to generate a magnetostatic equation for permanent magnets with model variables vars and model parameters pars.

reference.wolfram.com

 

[smowry]

So the first thing you might recognize is that based on my research, Borresen used a tweeter motor assembly in his woofer implementation. This is a major subtopic of this discussion.

 

[smowry]

By filing a US Patent Application, Borresen has fully disclosed his claimed invention, HUH! Borresen's US Patent will provide us with important information on his ironless woofer.

https://patentimages.storage.googleapis.com/ee/58/7f/2bcce6b865b6e7/US10993035.pdf

 

[smowry]

I am retired and I have not practiced in 10 years. So I don't have my software running. If there are folks out there that have non-linear electromagnetic FEA and could support this discussion that would be very helpful. The model is really simple 4 x box cross sections. NTK can you create a contour plot of the flux density inside the magnets (the cross sections) in the axial direction? Not the magnitude, the flux density, B in the axial direction. Can you plot the magnetic field strength, H within the magnets?

What BH curve should I use? My guess is N50. But what about the cryogenics? Forget that for now.

 

[amirm]

@NTK is our resident math expert and @René - Acculution.com , our expert in simulation of transducers.

 

[smowry]

Outstanding! And NTK is really fast.

I have asked Rene to join us.

Thank you Amir.

 

[smowry]

If no one is asking questions, then I will start with a question. Why is the STEALLUS woofer topology so different from the Almost air core tweeter motor topology?

Firstly, the woofer needs to be robust with reasonable power handling. While the tweeter is less so. This brings me to first concern regarding the Borresen implementation.

When there is no high permeability material (iron or steel) on the magnet surfaces in the direction of magnetization, the operating point of that magnet(s) can be very low.

It can be observed that as temperature increases, the BH curve becomes nonlinear and forms a knee further up the curve. If the load line falls below the knee at the respective temperature curve then permanent demag will occur. Various load lines are drawn with values from -0.2 to -2.0. The position of the load line is essentially controlled by the ratio of the magnet's surface to thickness. If the designer wants to raise the load and thus the operating point of the magnet, then he makes the magnet thicker relative to the surface. Then almost paradoxically, if he wants to increase the flux density in the gag, then he will increase the surface relative to the thickness.

Note that ring magnets will have somewhat lower operating points for the same surface to thickness ratios than a disk (no ID) due the losses within the ID. As the Energy Product of the magnet is increased, the knee will appear at a lower temp.

So then there's another almost paradoxical relationship between magnet's grade and the operating point positioning. However, Borresen claims cryogenics and that is still yet another paradox that I hope to discuss later on.

Warning: because all ironless motor assemblies contain magnets with different directions of magnetization (typically opposed) the magnetic assembly cannot be re-magnetized if demag occurs. The transducer must be reworked or scrapped.

Are there any questions, comments, and/or concerns so far?

 

[smowry]

If we were discussing the Purifi woofer, then Lars would be all over me. But where oh where is Elizabeth Borresen. "Where you at boy?"

 

[olieb]

Are there any questions, comments, and/or concerns so far?

I can not really make out the geometry of the Steallus motor.
Is that about right?

What is the stuff between the spiders?

 

[smowry]

No questions, then I will continue. The STEALLUS woofer has 2 x tall (very tall) NeFeB magnets. The why not use the Almost air core topology and just scale it for a woofer application? With the STEALLUS topology I can implement a large OD voice coil. If I try to implement a large OD voice coil, I tend to lose the outer magnet rings and in limit as the magnet separation increases, effectively the Almost air core topology begins to approach the STEALLUS topology. A woofer with no motor part in relation the voice coil OD will remove failure modes that relate to voice coil rubbing and will reduce manufacturing rejects, just so long as the voice coil ID and the PHALLUS annular clearance is gauged.

My first claim is that the Borresen ironless woofer implementation is controversial for the following reasons.

1. The magnets seem too thin and thus the operating points seems low, in other words, the motor is under designed.
2. I had investigated scaling the Almost air core topology in hopes of implementing a high performance woofer or even a midbase, I could not do it. I just could not find a robust solution that could provide the Beta targets I was looking for. Where Beta = (Bl)^2/Re (N^2/W).
3. Borresen used a tweeter motor for his woofer or is it a midbase?
4. My research indicated that any ironless woofer motor assembly needs between 10x and 20x the magnet volume (depending on topology) relative to a typical motor design with a high permeability magnetic circuit containing a gap.
5. The reverse roll surround is the worst choice available with regards to manufacturability and it reduces the "wheel base" (distance between surround and spider). This increases the likelihood of rocking modes. At BOSE reverse rolls were banned! Without question the surround topology that sets the standard for woofer surrounds is the Purifi surround. Having said that, I claim that the Borresen woofer surround topology is at the lowest industry standard and cryogenics cannot help him with this.
6. Where are the magnet ID and OD heatsinks? More cost containment related to tooling et al. perhaps.
7. The annular radius of the spider is small for a woofer and will limit linear displacement. Note that the STEALLUS has a much larger spiders in a complementary array and a huge wheel base. STEALLUS was designed for high displacement, whereas the Borresen woofer is designed for low displacement relatively speaking. The suspension is under designed.
8. The cone appears to be semi-press paper on the front side and non-pressed on the rear side but I could be mistaken. Myn choice wold be a high performance sandwich composite material. I claim that paper cones are obsolete. General Danny Richie touts paper cones.
9. The 180 degree lead dress is a good thing; however, for the highest reliability, I prefer 180 degree flat tinsel that is integrated into the spider and/or attached to the spider.
10. Those long fasteners concern me. They seem to violate a design for manufacturability and reliability rule. Avoid fasteners whenever possible.

Having said this, I cannot recommend the Borresen woofer; however, I can recommend the Purifi woofer as a high performance alternative.

 

[smowry]

olieb,


Yes that's correct but I used just 2 x tall opposing magnet. In early 2000's magnets from China were cheap. I should mention that tall magnet like that require a large and powerful magnetizer.

The only thing between the spiders is air. 2 x spiders with voice coil bonded to a plastic ring to form a module. That module is then bonded to the basket and PHALLUS subassembly using a gauge.

 

[olieb]

Yes that correct but I used just 2 x tall opposing magnet.

Yeah, makes sense. The section view just looked as if there were 4 magnets. (And it looked as if there was something between the spiders other than air.)

Why do these magnets have to be so tall?
a. linearity?
b. for a high enough working point?
Could one make them shorter and add some steel instead to cut cost? Or is the iron too bad, even so far "out"?

 

[smowry]

Why do these magnets have to be so tall?
a. linearity?
b. for a high enough working point?
Could one make them shorter and add some steel instead to cut cost? Or is the iron too bad, even so far "out"?

a. Linearity no, that relates to Bl(x) and Le(x,i).

b. Absolutely, I see setting the operating point high enough to have a robust driver as the primary design challenge. I hope we can get some models because I feel that it can be shown that the Borresen motor in not robust and is what I used to call wimpy back in the day.

c. That topology with steel end cap or what magnet folks call "keepers" is what I called STEALLUS X and yes you are correct.

 

[Svend P]

where oh where is Elizabeth Borresen

Elizabeth is a womans name. I think you mean Michael Børresen?

 

[smowry]

Sevend P,


Before I became an R&D engineer, I was a package driver for UPS for 17 years in Rhode Island (a truck driver). So I can tend to be a bit of a wise guy sometimes.

Going forward, I will refer to him as Elizabeth Michael Børresen, in honor of the biggest abuser of patents of this generation, Elizabeth Holmes. Without the support of the USPTO, Elizabeth Holmes would not have been able to perpetrate her campaign of fraud. Elizabeth Holmes claimed more than 200 inventions. So now Elizabeth Holmes lives in the gray bar hotel. More later but Elizabeth Michael Børresen has snake oil in his veins. We will get to that. Did Borresen invent the 4 x NdFeB ring magnet ironless motor assembly as he claimed in 2019? "You know he read it in a magazine." Sir Elton John. https://patentimages.storage.googleapis.com/ee/58/7f/2bcce6b865b6e7/US10993035.pdf

 

[DanielT]

I am reading the translated text below and trying as best I can, as a happy and curious amateur, to understand.So please do not ask me to clarify what I-or says because I have no knowledge about it. In any case, this thread is about what a developer of transducers has to say about magnets and steel. You may find it interesting.

Via Google Translate (I blame Google for any errors in the translation):

For my part, I am thinking more directly of eddy currents induced from the coil in the steel. The magnetic field in connection with these eddy currents counteracts the induced field and reduces the inductance. The eddy currents naturally also lead to resistive losses. The steel exhibits nonlinear hysteresis effects (memory effects), which lead to inductance-related distortion (the current driving the coil becomes distorted) when the magnetic field varies in strength and changes direction in the steel grains. These hysteresis effects can have time constants of minutes to microseconds and are very difficult to handle. You get a distortion floor in the reproduction, typically around 0.1 - 0.5% in the midrange, which remains even at low sound levels. The eddy currents correspond to very strong magnetic fields in the steel, which worsens the hysteresis distortion.

Short-circuiting rings around the magnetic gap can reduce the eddy currents in the steel, but do not eliminate them. Placing magnetic material close to the coil is an advantage because it avoids the sensitive steel in this area, but it also leads to lower maximum flux densities and thus lower voltage sensitivity for the speaker.

To truly eliminate the eddy currents, it is not possible to use sintered neodymium because the material has relatively low resistivity. The polymer type of neodymium magnets (bonded) in practice act as insulators and are preferable in such an application, but these only produce about half the flux density compared to sintered ones.

Here is a link to an interesting doctoral thesis that goes all out for a completely steel-free magnetic circuit (even the return circuit) built up of only "bonded magnets" with different magnetization directions and the distortion reduction is impressive (however, the voltage sensitivity is low because ferrite magnets are used here):https://tel.archives-ouvertes.fr/tel-01279478/document

Incidentally, saturation magnetization of steel is not a feasible path if one is looking to almost eliminate the eddy currents because in practice one does not get close enough to full saturation and therefore there is still the possibility for the eddy currents to arise (certainly in reduced form, but still strong enough to cause problems).

The most elegant way to approach the whole problem of inductance hysteresis distortion is to switch to current drive from the amplifier (high-impedance drive), but this of course leads to a lot of new difficulties.

faktiskt.io • Visa tråd - Ny 15” från Scan-Speak

 

[boxerfan88]

Could the "iron-less woofer" result in this sort of FR performance?

.